GPU Computing for Space Sciences

The rapid and incredible increase in computing capabilities of Graphics Processing Units (GPU) over the past few years, particularly through public applications such as video games, promotes the use of a GPU to do general purpose scientific and engineering computing, also known as 'GPU computing'.

Following several years of research and theoretical debates, led and pushed by the most important universities and GPU manufacturers, the time has come for GPU to address vertical markets such as chemistry, finance, medical and space sciences.

Thanks to its dedicated expertise on GPU computing and Space Sciences, Telespazio VEGA has the capabilities to master this new technology and address clients' specific requirements.

The need for GPU Computing

The model for GPU computing is to use a Central Processuing Unit (CPU) and GPU together in a heterogeneous computing model:

• The sequential part of the application runs on the CPU
• The computationally-intensive part runs on the GPU
• From the user’s point of view, the application runs faster because it is using the high-performance of the GPU to boost the performance

As shown in this figure, the few but 'very clever' CPU cores are running in collaboration with the numerous but 'less clever' GPU cores. From the theoretical point of view, for some operations, the speed-up proposes by GPU computing may be from a factor from x2 to x500 for the same operation realised either only CPU or CPU and GPU combined.

Being reasonably 100 times faster for a common operation necessarily opens new perspectives:

• From the user point of view, reducing the computing time means that he can go faster on ordinary processing and imagine new way for data processing and manipulation
• From the financial point of view, reducing the computing costs has an immediate impact on overall activities: the budget initially allocated to buy thousands of CPU may be invested in other R&D project or clever placement.

GPU Computing and Space Sciences

Space Sciences, including Earth Observation, is one of the most demanding activities in terms of computing power. Where as, on one hand, the HPC market proposes new solutions (combining various computing technologies), on the other hand, the everyday manufacturers (Dell, HP, etc.) can easily push the GPU computing market with the introduction, at very limited cost, of GPU processors directly inside the scientist workstation.

The time has come for the Space community to handle this new computing power directly and easily by introducing it into their everyday CPU computing activities.

GPU computing will shake up Space scientists, not only by accelerating by double digit factor their every day applications, but also by giving them new ways to handle their data. Additionally, GPU computing will address and revolutionise Space Sciences by the way it will easily speed-up very large images, and data processing and handling.

In a global context, where date age is a key factor to lead the market, GPU computing can make the data available faster and easier at a reasonable cost.

Telespazio VEGA and GPU Computing

Based on its unique recognised and combined knowledge both in the Space area (CNES, ESA) and in GPU computing (NVidia), Telespazio VEGA provides the best technological and tailored solution for the Space Science sector. Solutions will be based on:

• More than 30 years' expertise in Space Science with the development of Satellite Ground Segments, Satellite Training Centres, Earth Observation highly complex algorithms, etc.
• More than ten years' expertise on GPU programming from the original technology (OpenGL) to the latest ones (Cuda, OpenCL, etc.)

This unique combination of internationally recognised scientists and GPU computing expertise makes Telespazio VEGA your right and first partner to discover and make profit of GPU computing. We can analyse, port and optimise your current application to GPU computing making it available to your every day application.

Just as Telespazio VEGA recently did for the French National Space Agency (CNES), we can analyse, port and optimise your current application to GPU Computing making it available to your every day application.

Contact Telespazio VEGA for further information about GPU Computing